Firm-Specific Human Capital and
the Theory of the Firm
Margaret M. Blair
The Brookings Institution
First Draft: April 1996
Version of February, 1997
An earlier version of this paper was prepared for a conference on Employes and
Corporate Governance, Columbia University School of Law, Nov. 22, 1996. I
welcome comments and feedback. Please do not cite or quote without permission. I
am grateful to Gabriel Loeb for research assistance on this paper, to Bengt
Holmstrom, Victor Goldberg, Greg Dow, Louis Putterman, David Ellerman, and
participants in the Columbia Law School conference for comments and feedback, and
to the Alfred P. Sloan Foundation, Pfizer Inc., and an anonymous donor who provided
grants to the Brookings Institution to help support this research. The opinions
expressed in this paper, and all errors of fact and of judgment are those of the author
and are not to be attributed to Brookings, its officers or trustees, nor to any of those
who have helped to fund the research.
I. Introduction. The “theory of the firm” in modern economics has advanced
dramatically from the simple black-box production function model that is still
presented in most undergraduatelevel microeconomics textbooks. Likewise,
economic theories of the employment relationship have advanced far beyond simple
supply and demand curves for labor. But, while the nature of the relationship between
a firm and its employees would seem to be a central, perhaps defining, feature of the
firm itself, economists have generally studied questions about the nature of the firm,
as well as the ownership rights and governance structure of firms, separately from
questions about the structure and terms of employment relationships, and they have
not yet produced a unified theory of the firm that adequately explains and accounts for
the role of “human capital.”
1
Of course, there may be no single, simple theory that explains the numerous complex
organizational forms and relationships that we actually observe. But a growing body
of theoretical and empirical work, by scholars working from a number of different
angles, suggests that specialized investments -- investments whose value in a
particular enterprise greatly exceeds their value in alternative uses -- play a critical
role in determining the boundaries of firms, and the allocation of “property” rights
(i.e., risks, rewards, and control rights) within firms. For example, a number of
different scholars have stressed the incentive benefits that flow from assigning control
rights over the assets of a firm, or over the firm itself, to parties who make important
firm-specific investments of one sort or another. But much of this literature has
focussed on investments in physical or other alienable capital. There are two
exceptions to this general rule, both of which will be discussed briefly later in 1 this
paper. The first exception is the small body of theory and empirical research on labormanaged firms (See e.g., Ward (1958), Domar (1966), Vanek (1970, 1977), Meade
(1972), Furobotn and Pejovich (1974), Putterman (1984), Ellerman (1986) and Dow
(1993). The second is the effort by Masahiko Aoki to develop a “cooperative game
theory of the firm.” See Aoki (1984). Neither of these strands of the literature have
had much impact on mainstream economic thought.
2
To be sure, scholars have recognized and discussed the importance of firm-specific
investments by employees for decades. Specialized knowledge and skills of
employees are increasingly understood to be important factors that influence the
structure and character of the employment relationship, and the allocation of risks and
rewards between employees and employers.
But work by labor theorists has generally argued that investments in firm-specific
human capital are protected through such institutional arrangements as collective
bargaining or promotion ladders -- arrangements that do not normally convey control
rights over the firm itself. In the models used by these theorists, the contributors of
firm-specific human capital are generally viewed as parties to transactions with the
firm (where the firm is apparently some well-defined entity on the other side of the
relationship) rather than parties to the firm itself. Meanwhile, the modeling approach
that has dominated legal and normative discussions of corporate governance questions
in recent years has been the principal-agent model, in which managers are viewed as
agents of shareholders. Principal-agent analysis can be a misleading tool for analyzing
certain kinds of corporate governance problems for at least two reasons, however.
First, the canonical principal-agent model is asymmetric and hierarchical. It assumes
that we know who the principal is and who the agent is in a given relationship, and it
focusses on structuring the relationship in such a way that the agent will be motivated
to do what the principal wants. It ignores any problems of enforcement in the other
direction. That is, it assumes away all of the potential problems that might arise in
getting the principal to deliver on its end of the bargain.
Second, the model assumes that the principal is a well-defined entity (usually an
individual) with an unambiguous agenda. The principal knows what it wants the
agent to do. Thus principalagent analysis doesn't help to sort out the complexities that
arise when there are multiple parties involved in a joint enterprise, perhaps with many
different goals.
3
These two limitations suggest that, while principal-agent analysis has provided some
important insights into certain kinds of contracting problems, it tells us very little
about the essential nature of corporations. Moreover, it can be very misleading when
used to draw normative implications about corporate governance if used in ways that
ignore, or mask, the underlying ambiguity about who the principal is in a given
context (Is it the “corporation”? Is it the shareholders? Is it management? Are any of
these necessarily a monolithic entity with an unambiguous agenda?), or how the goals
of the principal are determined.
A small, but important and growing body of work on the economic theory of the firm,
however, attempts to explain, or at least explore, the nature of the firm itself. Why are
some productive activities organized within firms and others not? What is the
difference between a within-firm relationship, and one that is governed by contract, or
across markets? What determines who is in the firm and who is out? This work
generally considers the institutional arrangements for encouraging and protecting
relationship-specific investments, and/or for encouraging and rewarding effort toward
a group goal by individual employees, as defining features of the firm itself. Although
most of these theoretical efforts are new, and still quite sketchy, they may well lead to
normative implications about the allocation of claims and control rights in publiclytraded firms that are different from those drawn from the principal-agent models that
have been so prevalent in legal analysis in the last two decades.
In this paper, I will assess the status of theoretical work on the role of firm-specific
human capital in the theory of the firm, and comment on the implications of some of
the newest work for legal analyses of control rights and of the responsibilities of
managers in firms.
II. A Brief History of Thought
4
Firm-specific (or relationship-specific) investments have been important in two
subfields of economics, labor theory, and the theory of the firm.
A. Labor Theory.
Gary Becker (1964) coined the phrase “human capital” to refer to the fact that much of
the skills and knowledge required to do a job could only be acquired if some
“investment” was made in time and resources. In his pathbreaking work on human
capital, Becker considered the implications of the fact that some of the knowledge and
skills acquired by employees have a much higher value in a given employment
relationship than they do in other potential relationships. This fact, he speculated,
would influence choices about wages, investments in training, and other terms of the
employment relationship, Becker argued that such specialized knowledge and skills
may often be productivity enhancing, and are therefore likely to be an important part
of the employment relationship in practice. But, he noted, they introduce a
complication into simple models of wage determination and equilibrium employment
levels. In particular, the labor services of employees with specialized skills can no
longer be modeled as undifferentiated, generic inputs, for which equilibrium price
(wages) and quantity (number of employees or number of hours of work) are
determined by the intersection of supply and demand curves. Once employees are
understood to have specialized skills, it matters which employee does what job for
what firm.
“If a firm had paid for the specific training of a worker who quit to take another job,
its capital expenditure would be partly wasted, for no further return could be collected.
Likewise, a worker fired after he had paid for specific training would be unable to
collect any further return and would also suffer a capital loss,” Becker noted. Where
investments in specific skills are important, Becker reasoned, it is no longer a matter
of indifference
5
“whether a firm's labor force always contained the same persons or a rapidly changing
group.”
2
Thus, although Becker's primary purpose was to study and explain the economic
incentives for investments in training and education, along the way he introduced a
concept that provides an important rationale for long-term relationships between firms
and their employees. Doeringer and Piore (1971) built on this insight to develop their
theory of internal labor markets. They argued that investments by firms in specialized
training encourage firms to put in place other institutional arrangements designed to
stabilize employment and reduce turnover. The organizational stability that results
from these practices, in turn, facilitates further development of specific skills.
Doeringer and Piore further argued that the use of mass-production technology, with
its detailed division of labor, required specialized skills and made stable employment
relationships more important. 3
Becker also argued that employees and employers would be likely to split both the
costs and returns from specialized training, in order to provide an incentive for both
parties to stay in the relationship. One of the implications of his arguments was that
employees would typically earn less
4
See Becker (1964), p. 21.
2
Jacoby (1990) questions this conclusion. Though he concedes that empirical
evidence supports 3 a shift from the late 1800s to at least the mid-1970s toward
greater job stability, but he argues that “there is little evidence that the shift resulted
from a growing reliance on firm-specific techniques or skills. In fact, the evidence
suggests that the opposite was true: that technology and job skills became less, rather
than more firm-specific over time.” See p. 323. The notion of firmspecific skills that
lies behind Jacoby's assertion is that of skills necessary to carry out unusual or
specialized production techniques. It does not encompass the admittedly less welldefined notions of knowledge and relationships necessary to participate effectively in
a given organization.
Hashimoto (1981) subsequently provided a formal model that suggested that the
division of the 4 costs and returns from training would be split according to a formula
that was a function of the relative probabilities of layoffs versus quits, and the costs of
evaluating and agreeing on both the worker's productivity in the firm and his
opportunity cost, or potential productivity in an alternative firm.
6
than their opportunity cost during the early stages of their employment relationship
(while they were in training, for example), and more than their opportunity cost later
in their relationship. An earnings pattern like this would produce an “upward sloping
wage-tenure profile,” an empirical regularity that labor economists before Becker had
observed, and that work by subsequent scholars has documented extensively.
Consistent with the “firm-specific human capital” hypothesis, labor 5 economists have
also observed that long-tenured employees typically earn quite a bit more than their
short-run opportunity cost. This fact is confirmed through studies of layoffs, which
show that longtenured employees laid off through no fault of their own (as a result of
plant closings, for example) typically earn 15 to 25 percent less on their next jobs.
These estimates and others by scholars doing 6 related work suggest that the aggregate
returns to investments in firm-specific human capital could represent as much as 10
percent of the total wage bill of the corporate sector -- a figure that is of the same
order of magnitude as all of corporate profits.
7
Not all labor economists are convinced that the empirical evidence that wages rise
with tenure, and that wages of long-tenured employees often exceed short-run
opportunity costs, should be taken as evidence that employees have acquired
substantial amounts of firm-specific human capital. Some labor economists have
argued that other features of the labor market could account for these empirical
regularities. The other explanations that have been offered also imply that labor For
recent contributions to this literature, see Topel (1990), Topel (1991).
5
See, e.g., Jacobson, LaLonde, and Sullivan (1993).
6
See Blair (1996), p. 10, and footnote 4.
7
markets would exhibit involuntary unemployment. Hence they have been very
important in the debate about whether labor markets clear.
8
But these stories (and the empirical studies that support them) do not generally rule
out the possibility that firm-specific human capital is an important factor in
determining the structure of many employment relationships. Indeed, most labor
economists believe such investments are important in many situations.
9
Once acquired, knowledge and skills that are specialized to a given enterprise are
assets that are at risk in that enterprise, and as such, they inevitably present a
contracting problem for the employee and the firm. If the firm compensates the
employee up front and fully for the costs of developing and using such assets, the
employee could, in principle, take the compensation and the assets and walk out the
door. Suppose, however, that the firm does not fully compensate the employee up
front, but instead, compensates him or her in the way that Becker predicted, with a
lower wage at first, and a promise of a higher wage later. That employee would then
have a stake in the firm that is unrecoverable except as payments are made to the
employee out of the economic
For a summary of arguments on the efficiency of
non-market clearing wages, see Krueger and 8 Summers (1988) and Weiss (1990);
for evidence on non-market clearing wages and employment practices, see Katz and
Summers (1989) and Dickens and Lang (1993).
As the reader may have guessed, my interest in firm-specific human capital is
derived from my 9 primary interest the effect of various corporate governance
arrangements on the incentives and risks facing employees of corporations.
Characterizing the problem as one associated with investments in human capital
emphasizes the parallels to the corporate governance problems facing investors in
equity capital. The governance issues raised by investments in human capital are
similar, however, to those that arise with any compensation scheme that is “backloaded” for whatever reason, as well as with any compensation scheme that results in
employees being paid more than their (short-run) opportunity cost. Matching models
predict back-loaded compensation schemes, and both matching models and efficiency
wage models predict wages in excess of short-run opportunity costs. In such cases,
employees have something of substantial value at risk in the firm that can only be
recovered over time, and that can be expropriated, or that can be lost altogether if the
employees lose their job with their current employer.
8
surplus generated by the relationship. This stake is generally regarded as very difficult
to protect by means of explicit contracts: The firm can't enforce a contract that
requires the employee to stay and utilize those skills in the firm. And, because the
skills in question are likely to be hard to define, let alone measure, the employee can't
enforce a contract that requires the firm to pay for the development and use of skills.
Yet, without explicit contractual protection, the stakes held by providers of firmspecific human capital take on many of the characteristics of the stakes held by
providers of equity capital.
B. Theory of the Firm.
The earliest efforts by economists to get inside the black-box production function
model of the firm focussed on the fact that production in firms is characterized groups
of people working together, and organized hierarchically, with some individuals
having the authority to make decisions about how people and resources are used. The
authority relationship, Coase argued, is substituted 10 for a series of market
transactions, because, for a variety of different reasons, the central authority figure in
the relationship can coordinate activities more efficiently than individual input
providers could if they were all contracting with each other separately. From this
initial insight, economists took the theory of the firm in two different directions. One
direction stressed the importance of joint production technologies, in which it is
difficult to measure and reward the individual productivity of team members. In this
approach, the purpose of 11 “If a workman moves from department Y to department
X, he does [so] . . . because he is 10 ordered to do so,” writes Coase. “Outside the
firm, price movements direct production, which is coordinated through a series of
exchange transactions on the market. Within a firm, these market transactions are
eliminated and in place of the complicated market structure with exchange
transactions is substituted the entrepreneur-co-ordinator, who directs production.”
See Coase (1937), p. 19.
See Alchian and Demsetz (1972).
11
9
the firm is to provide a mechanism by which a central authority can monitor the
activities of the team members to be sure they all carry their weight. The central
authority has the incentive to monitor effectively because he gets to keep any surplus,
over and above what he has to pay the team members. The principal-agent models
can be understood as extensions or supplements to this key idea. The second approach
has been to focus on circumstances in which it might be less costly to organize
production within a firm. This second approach has stressed, among other things, the
importance of investments in specialized assets. I will review the literature on the
latter approach first, and then return to the problem of team production.
1. Transactions costs theory. Oliver Williamson has identified several features of
transactions that make it costly to trade in impersonal, arms-length markets. Where
these features apply, transacting parties might choose to administer such transactions
through hierarchical governance arrangements. These features include the longevity
and “asset-specificity” of 12 investments made to engage in a given transaction. The
former means that the asset will generate its return over time, while the latter imposes
limits on the ability of the transacting parties to redeploy assets to other uses if for any
reason they are unhappy with the returns they are getting in the given enterprise.
Other features that encourage hierarchical administration rather than market
transactions, according to Williamson, include the uncertainty and complexity of the
transaction (a
problem which is exacerbated when assets are long-lived); the “bounded rationality”
of the
See Williamson (1975, 1985). “Markets” and “hierarchies” were treated as
dichotomous 12 organizational forms in Williamson's early work, but more recent
work, by Williamson and others, tends to view them as two ends of a continuum of
possible structures in which transactions take place, with various kinds of contracts in
between. See e.g. Williamson and Bercovitz (1996), Baker, Gibbons, and Murphy
(1996), and Kreps (1996).
10 transacting parties (which makes it impossible for them to anticipate all possible
outcomes and complications and to write “complete” contracts that specify what is to
happen under each scenario); and the tendency of transacting parties to be
“opportunistic.” Of these, asset specificity is central in Williamson's theory of the
firm. Assets that are not specific to a given enterprise or transaction can be readily
redeployed and hence are not at risk in a given relationship. But when assets are
specific, they are, almost by definition, at risk, and the other problems (uncertainty
and complexity, bounded rationality, and opportunism) become important.
Williamson's work spawned a literature on the contracting problems that arise when
assets are specific. Klein, Crawford and Alchian (1978), for example, argued that
when two contracting parties each make investments that are specific to their
relationship, either party can attempt to expropriate the returns from those investments
by threatening to “hold up” the enterprise. The potential “hold-up” problem, they
speculated, would encourage the contracting parties to integrate their operations
vertically (the supplier would acquire the customer, or vice versa) so that all of the
assets that were specific to the enterprise would be owned by the same party. For
example, if one party owns a coal mine, and the other party owns a power plant built
at the mouth of the coal mine and designed to use coal from the mine, the two parties
might find themselves in frequent disputes about the price and terms on which the
coal is to be sold to the power plant. But if a single party owns both the mine and the
power plant, and administers them jointly, this owner would probably try to maximize
the joint return, and would no longer be tempted to waste resources in haggling over
the terms of trade between the two units.
Empirical research that has attempted to test the Williamson and Klein, Crawford, and
Alchian hypotheses has generally confirmed that firm-specific investments are
important in determining ownership structure and degree of vertical integration. But
this research has taken an 11 interesting twist. Monteverde and Teece, for example,
studied parts production in the automobile industry to ask under what circumstances
firms might choose to undertake production in-house rather than contracting
production out to a supplier. They argue that vertical integration might not be
necessary if the specialized assets used in production of the parts include only physical
capital, such as tools or dies. The hold-up problem, in this case, can be avoided if the
automobile assembly company owns the specialized tools and leases them to the
contractor who produces the parts. Such arrangements, which Monteverde and Teece
refer to as “quasi-integration,” are commonly observed in the auto parts production
business. But where the specialized investment involved in producing the parts is in
nonpatentable know-how and skills, Monteverde and Teece argue that
quasiintegration will no longer solve the hold-up problem. They speculate that full
integration will be required to minimize the transactions costs. Consistent with this
hypothesis, they found in one
13
study that quasi-integration arrangements are associated with two indicators of the
extent of investment in specialized tools and equipment, and, in another study, that
full integration is associated both with an indicator of the extent of specialized knowhow involved in developing and engineering the component, and with an indicator of
the extent to which the component's design
Wiggins distinguishes transactions
and relationships that are “integrated” from those that aren't 13 by noting that the
former are arrangements “where both monitors are compensated out of a common
residual income stream.” (Wiggins, 1991, p. 615.) In other words, to be “in” the firm,
rather than to be on the outside contracting with the firm, means to be sharing in a
“common residual income stream.” The notion of a common residual income stream
only makes sense if there are cospecialized assets generating that common income
stream. Wiggins notes that, with a given set of demand and supply functions for
participating parties, simple long-term contracts can be devised that can give both
parties to a transaction incentives to make the optimal ex ante transaction-specific
investments. But, he says, “the compensation arrangements that solve the effort
problem undermine the parties' ability to adjust output under the contract to reflect
cost shocks.” Integration, he claims, can overcome “the incentive to distort shocks by
compensating both monitors out of the common residual of the integrated firm.” This
comes, of course at a cost of some of the intensity of the incentives facing both
parties.(p. 617)
12 affects the performance and packaging of other components, a feature they refer to
as the component's “system effects.” Similarly, Masten, Meehan, and Snyder (1989)
find that, in 14 regressions in which both investments in specialized knowledge and
investments in specialized equipment are used to explain vertical integration,
investments in specialized knowledge have much more explanatory power. 15 Neither
Williamson, nor Klein, Crawford, and Alchian offer much insight into how it is that
organizing production within a firm solves the hold-up problem associated with firmspecific investments by multiple parties, nor which of several participants in an
enterprise should be the “owner” of an integrated enterprise. Grossman and Hart
(1986) address this gap. They argue that organizing production in a firm solves the
incomplete contracting problems by assigning to one party -- the “owner” -- all of the
“residual” control rights over the use of the assets in the firm – those rights that are
not specifically contracted away to other participants. Under the terms of the
Grossman and Hart theory, firms are defined as bundles of assets under common
ownership, where ownership implies control over the use and disposition of the assets.
16
See Monteverde and Teece (1982a, and 1982b).
14
See Masten, Meehan and Snyder (1989).
15
Grossman and Hart's arguments have been called the “property rights theory of
the firm,” but, 16 even though the first sentence of their 1986 paper asks “What is a
firm?” (See p. 691.), their story is not so much a theory of the firm, but a theory of
role played by ownership rights in solving complex contracting problems. Hart and
Moore (1990) show how ownership rights over a firm's physical assets -- specifically,
the ability to exclude others from the use of the assets -- enhance the ability of the
owner to control the workers who work with those assets. If the workers require
access to the assets to be productive, they will have an incentive to do what the
"owners" of those assets want them to do. Thus Hart and Moore add to the notion of
ownership a theory about what it means to be in or out of the firm (an “employee” or a
“contractor”). 13
Grossman and Hart's model considers a situation in which participants in an enterprise
must make firm-specific investments that are very difficult or impossible to contract
over. Their model leads to the conclusion that the ownership rights in the firm should
go to the party whose firmspecific investments add the most value to the enterprise,
but are the most difficult or impossible to contract over. Ownership rights ensure the
party who must make these investments that her claim to a share in the rents generated
by the investments will not be expropriated ex post by the other participants. Hence,
they help ensure that she will have the ex ante incentive to invest optimally.
Williamson's early work acknowledged that human capital, as well as physical capital,
can be transaction-specific. In later work, he identifies some features of
organizations, such as team accommodations, informal process innovations, and
knowledge of codes and procedures, that tend to make incumbents more valuable to
employers than workers hired on spot markets might be. And, 17 he argues that
transactions in which investments in specific human capital are important must
include some sort of safeguard for those investments, noting in particular that in such
transactions, “continuity between firm and worker are valued.” Firm-specific human
capital, he says, must be “imbedded in a protective governance structure lest
productive values be sacrificed if the employment relation is unwittingly severed.”
18
Williamson himself devotes little intellectual energy to identifying, or assessing the
safeguards or protective governance arrangements that, according to his theory, should
exist in relationships between employees and firms that employ them. He mentions
severance pay, and job security (of the kind commonly thought to exist for employees
of Japanese corporations) in passing, as mechanisms for protecting worker
investments in firm-specific skills, and he mentions pensions as a mechanism for
providing incentives that discourage employees with specialized skills from quitting.
He also argues that collective bargaining through unions and “internal governance 19
structures” (such as job ladders, grievance procedures, and pay scales) can serve
efficiency purposes by providing a protective governance structure for idiosyncratic
investments in skills by workers.
20
But Williamson's work on these issues follows the tradition of labor theorists (see
discussion above) in treating employees as contracting with the firm, where the firm is
understood to be a well-defined entity on the other side of that contract.
21
See Williamson (1985), pp. 246-247.
19
See Williamson (1985), pp. 254-256. See also Williamson, Wachter and Harris
(1975).
20
Williamson also notes that when the intensity of firm-specific human capital is high,
unions may actually be inefficient mechanisms for protecting that human capital.
"Among other things, a single union operating under a uniform agreement will have
difficulty aggregating the preferences of a disparate membership. To negotiate
discriminating terms is at variance, however, with the egalitarian purposes of unions."
See Williamson (1985), p. 265.
See also Epstein (1985) for a discussion of the contracting inefficiencies supposedly
caused by unionization under protective regulation. Epstein believes unions would
not survive without protective legislation, because, he claims, "at will" contracts are
so much more efficient. Nonetheless, Epstein concedes that "at will" contracts will not
suffice in situations where "one party to the employment relationship has to make a
very substantial capital outlay at the outset," or in situations where "the difference
between the contract wage and the market value of the services [is] very large." (See
p. 138.) Of course, significant investments in firm-specific human capital lead to
exactly these circumstances.
In fact, Williamson apparently assumes either that the protections that are
available to
21
employees for their firm-specific investments, whether explicitly contractual or
institutional, are fully adequate to protect them, or possibly that those investments are
of less significance than the investments made by those who contribute equity capital.
This assumption is implicit, for example, in his analysis of the transactions cost
benefits of corporate governance arrangements that give the right to elect board
members to shareholders, and not to other constituents: "Stockholders as a group bear
a unique relation to the firm," he asserts. "They are the only voluntary constituency
whose relation with the corporation does not come up for periodic review. . . .
Stockholders. . . invest for the life of the firm, and their claims are located at the end
of the queue should liquidation occur." See Williamson (1985), p. 304-305.
This assertion ignores the fact that, since investments in firm-specific human capital
can't be
15
2. Team production. Alchian and Demsetz argue that the role of the firm
is to organize team production, which they characterize as “production in which 1)
several types of
resources are used and 2) the product is not a sum of separable outputs of each
cooperating resource,
. . . [and] 3) not all resources used in team production belong to one person.” The
problem raised
22
by team production, according to Alchian and Demsetz, is one of metering output
(where the output
of any one individual is not separable from the output of his or her teammates) and
issuing rewards
in ways that will motivate team members to exert effort. Advantages arise from team
production if
the team members can accomplish substantially more by working together than they
could
accomplish by working separately. If this extra productivity exceeds the cost of
monitoring and
motivating team members to exert effort, then team production will be chosen over
individual
production methods.
Alchian and Demsetz go on to argue that, where team production is preferred, the
metering
and reward problem can be solved by having one team member specialize in
monitoring, and by
giving that individual both the authority to hire and fire team members, and a claim on
the earnings
from the enterprise, (net of payments to providers of other inputs, who are assumed to
be paid their
redeployed (by definition), they too are invested for life. Employees can deprive the
firm of their use, but they cannot get any benefit from doing so since the skills have no
value anywhere else. Hence, it is not clear how much protection is provided even by a
continuing opportunity for renegotiation. Moreover, employees have no claim at all,
except for past wages, if the firm is liquidated. Although we do not have precise
estimates of the aggregate value of investments in firmspecific human capital, there
are reasons to believe that it is large, and possibly of the same order of magnitude as
the aggregate value of equity capital. See footnote 6 and associated text above.
Williamson's dismissal of the importance of firm specific human capital for corporate
governance has, nonetheless, been echoed by numerous legal scholars. See, e.g.,
Roberta Romano (1996), at p. 3.
See Alchian and Demsetz (1972), p. 779.
22
16
opportunity cost). Thus they claim that their story provides an explanation for
capitalist ownership and control of firms.
Alchian and Demsetz's story focusses attention on the contracting relationships among
the participants within firms, and provides an explanation for hierarchical structures.
From that beginning, numerous other authors have addressed the contracting problem
between so called “principals” (the entrepreneur or central, capitalist authority figure
in Alchian and Demsetz's story) and “agents.” 3. Principal-agent theory. In principalagent models, employees are viewed as agents of the firm, and the managers of firms
are viewed as agents of the shareholders. The contractual problem is to design the
terms of the relationship in a way that will encourage the agents to make decisions and
otherwise behave in ways that benefit the principals. Jensen and Meckling (1976)
introduced the principal-agent approach to the theory of the firm in their classic
article in which the firm was viewed as a contracting mechanism between providers of
equity capital (the principals) and managers (the agents) designed to minimize the
agency costs of this relationship.
23
In recent years, principal-agent models have been very influential in studies of
corporate governance arrangements and performance, and have formed the basis of a
huge literature on the “market for corporate control,” and on incentive compensation
structures in firms. Principal-agent theory has also been influential in research on
labor relations in firms.
Jensen and Meckling argued that organizations “are simply legal fictions which
serve as a 23 nexus for a set of contracting relationships among individuals.” (p. 310.
Emphasis in original.) Hence they are generally credited as the source of the view of
the firm as a “nexus of contracts.”
17
The canonical principal-agent problem involves a transaction between two parties, one
of
whom must take an action that affects the other. For some reason, however, the
principal cannot
compensate the agent directly for the action itself. This may be because the action
itself is not
observable to the principal, or because the principal does not have the information or
knowledge
necessary to evaluate the action itself. Another complication is that the consequences
or observable
output of the agent's action is not a determinate function of the action. Otherwise, it
would be
possible for the principal to infer the action taken by observing the consequences.
Instead, in the
principal-agent problem, the output is assumed to be a stochastic function of the
agent's action,
and/or it is assumed to be measured with error.
Since the principal can't pay the agent for the action, he or she must devise a way to
pay for
the observable output or consequences of the action. The problem for the principal,
then, is to set the
fee schedule in a way that gives the agent incentives to choose actions that benefit the
principal.
If the agent is risk neutral, the agent can be induced to choose an optimal action if he
is
awarded the full payoff from the action. But if the agent is risk-averse, then this
arrangement is
24
generally not optimal, and the optimal incentive structure will be some function of the
payoff in
which the agent and the principal share the risks. The sharing of risks gives a riskaverse agent
higher utility for a given expected value of the payoff, but it inevitably reduces the
effectiveness of
the incentives provided by the fee schedule. Other devices that can be used to induce
effort by the
agent include flat fees, accompanied by a threat of termination if the agent is caught
shirking.
Risk-sharing fee schedules have been discussed extensively in the literature on
incentive
compensation systems for corporate executives, while flat fees accompanied by threat
of termination
This solution, of course, depends on the payoff being transferrable from the
principal to the
24
agent.
18
form the basis for both the literature on the market for corporate control as a
mechanism for
inducing managerial effort, and for the “efficiency wage” stories in labor theory.
25
But while principal-agent models have been useful in delineating certain kinds of
contracting
problems, they do not go to the essence of the nature of the firm itself. In fact, they
just assume the
existence of the principal, or the employer -- the entity on the other side of the
contract with management or labor.
4. Team production with long-term contracts. By itself, the Alchian and Demsetz
story offers no particular reason why the membership of the team couldn't change
from day to day, or hour to hour. In fact, they claim, “long-term contracts between
employer and employee are not the essence of the organization we call a firm.” They
argue, rather, that the relationship between firm and employee is equivalent to a series
of short-term contracts: “the employer is continually involved in renegotiation of
contracts on terms that must be acceptable to both parties.” 26 But we know that
fairly long-term relationships are the norm in large corporations rather than the
exception. So their story seems, at best, incomplete, for explaining the actual way that
large corporations typically operate. The Alchian and Demsetz story can be improved
by allowing for investments in firm-specific human capital, or other factors which
might make it advantageous to keep a particular team working together. Demsetz
himself has moved in this direction in more recent work. An important aspect of the
“nexus of contracts,” that makes up a firm, he says,
“Efficiency wages” refer to a
class of arguments in labor theory in which employees are 25 induced to perform well
by making it costly for them if they get laid off -- generally by paying them more than
their opportunity cost. See discussion above of non-market clearing wages. See also
Weiss (1990) for a thorough review of the efficiency wage literature.
See Alchian and Demsetz (1972), p. 777.
26
19
“is the expected length of association between the same input owners. Do the
contractual agreements entered into contemplate mainly transitory, short-term
association, which in the extreme would be characterized by spot market exchanges,
or do these agreements contemplate a high probability of continuing association
between the same parties. The firm viewed as team production exhibits significant
reassociation of the same input owners.”
27
Demsetz, in fact, goes on to define a firm as “a bundle of commitments to technology,
personnel, and methods, all contained and constrained by an insulating layer of
information that is specific to the firm, and this bundle cannot be altered or imitated
easily or quickly.” 28 In the types of team production that are characteristic of real
world firms, then, the human capital of the team members is worth more when applied
together with the human capital of the other team members than it is when applied
alone. With specific human capital, the productivity of a particular individual
depends not just on being part of a team, but on being part of a particular team
engaged in a particular task. But if it matters who is on the team, this complicates the
Alchian and Demsetz story because it is no longer clear that team members who
invest in specialized skills and who know they are especially valuable when deployed
with this particular team will be willing to accept only their (short-run) opportunity
cost in wages. Hence it is no longer obvious that the monitor will be able to collect all
of the economic surplus (the rents and quasirents) from the enterprise.
In some ways, the problem raised by investments in firm-specific human capital is
analogous to the principal-agent problem. The employee must take some action (e.g.,
acquire some skills, accumulate some special knowledge, develop some special
relationships with co-workers) which the firm cannot directly measure and for which
it cannot directly compensate the employee. The firm
Harold Demsetz, “The
Theory of the Firm Revisited,” in Oliver Williamson and Sidney
27
Winter, eds., The Nature of the Firm, Oxford University Press, 1991, pp. 170.
Demsetz (1991), p. 16.
28
20
can only observe (perhaps imperfectly) the outcomes of such investments. Wiggins
stresses the 29 similarities between the principal-agent problem and the problem of
firm-specific investments by noting that, whenever “one party performs first, he
effectively makes an investment specific to the trading relationship; he invests in a
specific asset. After investment, he relies on the other party to perform. The problem
is that the second party can only make limited commitments to follow through.”
30
But the firm-specific investment problem is different from the canonical principalagent problem in a critical way: it is symmetric. With firm-specific investments,
actions of both parties can affect the payoff from the investment. In the canonical
principal-agent problem, there is an implicit assumption that, once the fee schedule
has been determined, the actions of the principal have no further effect on the outcome
of the variable to which the fee schedule is tied. The outcome is realized, the fees that
were promised to the agent are determined as a function of that outcome, and the
agents are promptly paid. 31
In the firm-specific human capital story, by contrast, the employee takes an action that
affects the payoff for the firm, but the firm, in turn, can take actions that not only
affect the fee that the employee gets, but that affect the stream of rents and quasi-rents
generated by that action. The “The employment relation in general is one in which
effort and ability acquired through
29 training and self-improvement are hard to
observe,” notes Kenneth Arrow. “In one view, firms exist as a means of measuring
effort.” See Arrow (1985), p. 39.
Wiggins goes on to suggest that firms, contracts, and government regulations are,
or can be, 30 alternative mechanisms for solving these problems. See Wiggins
(1991), p. 604. Wiggins notes that agency theory implicitly “makes a strong
assumption about the credibility 31 of these two parties [the employer and the
employee], and this assumption drives all the results. The general contracting
problem, however, is to specify how promises are enforced, and then derive the
structure of the contract in a consistent manner from that specification.” See Wiggins
(1991), p.
646-647. firm can decide to close the plant where the employee works, and suddenly
there is no opportunity for the payoff to be realized, for example. Or the shareholders
can sell the firm to someone else who can fire the manager or dismantle the firm. 32
Hence, if firm-specific human capital is an important input in corporate enterprises,
the classical principal-agent model may be too one-sided to be adequate to the task of
describing the fundamental features of the employment relationship, and of the firm
itself.
Becker (1964) was clearly aware of this problem. See pp. ??. The point is
also close to that made by Shleifer and Summers (1988) in their critique of hostile
takeovers. Although Shleifer and Summers do not appeal explicitly to investments in
firm-specific human capital as the basis for the implicit contracts that they argue are
breached by takeovers, such investments would be one explanation of the quasi-rents
that are supposedly up for grabs in their story.
22
III. Contracting problems raised by firm-specific human capital: The potential
“holdup” problem presented by firm-specific investments of all kinds was discussed at
some length above. The hold-up problem makes such investments risky. But
investments in firm-specific human capital present other contracting problems too.
A. Risk-sharing. Firm-specific human capital is subject to two kinds of risks, the
risks that the rents and quasirents it generates will be expropriated (the hold-up
problem), and the risks that its actual value (its ability to generate rents and quasirents) will fluctuate. The risks to actual value can be further subdivided into the risks
that the particular skills will no longer be as useful in a given firm, and the risks that
the firm itself will no longer generate as much in total rents. Finally, the firm's risk
can be subdivided into idiosyncratic risks and systemic risks.
Much has been made of the idea that the corporate form facilitates a division of labor
in which managers specialize in decision-making, and outside investors specialize in
risk-bearing.
33
This approach to thinking about the nature of the firm and the contracting problems in
the firm, however, essentially ignores the risks borne by employees with firm-specific
human capital. At least two of the categories of risk facing employees are risks that
outside shareholders could not, even in principle, protect employees from. These are
the portion of the risks to underlying actual value that are due to systemic risks, and
the risks of expropriation, which are inherent in the contracting problem between
employees and firms. Numerous scholars have argued that employees are protected
from the risks of expropriation by the fact that the firm must be concerned about its
reputation for fairness. A good reputation
See Fama and Jensen (1983) for the
classic article that makes this argument.
33
23
enables it to contract on favorable terms with other employees in the future. In other
words, the 34 firm is assumed to have a longer time horizon than the employees. But
for the firm to play this role in the relationship, for it to be an effective bearer of
reputation, the identity of the firm must be reasonably stable. While corporations
have the potential of perpetual existence under the law, in recent years, corporations in
the U.S. have undergone dramatic structural and identity changes at a very rapid pace.
It seems likely that these changes have undermined the effectiveness of reputation as a
mechanism for ensuring employees of most firms that the firms value their reputation,
and will not renege on their end of the deal.
In practice, employees are also not given significant protections against the risks of
decline in the actual long-term value (rent-generating capacity) of their firm-specific
skills, and it is difficult to imagine how such protection could be provided in any
company over an extended period of time. But if employees share in the valuation
risks, this inevitably increases the expropriation risk 35
Milgrom and Roberts note, for example, that “the value of a good reputation
increases with the number of times it may be used. Thus, if one party to a relationship
has a longer horizon than another, the first party, when put in a position of power, will
have a stronger incentive to build and maintain a reputation . . . for using power well.
This argues for giving the residual power to whichever party to a relationship has the
longer horizon.” See Milgrom and Roberts (1992), p. 331. See Kreps (1990) for a
general discussion of the role of norms, reputation, and corporate culture. Finally, see
also Williamson (1985), p. 261: “Employers who have a reputation for exploiting
incumbent employees will not thereafter be able to induce new employees to accept
employment on the same terms. A wage premium may have to be paid; or tasks may
have to be redefined to eliminate the transaction-specific features; or contractual
guarantees against future abuses may have to be granted. In consideration of those
possibilities, the strategy of exploiting the specific investments of incumbent
employees is effectively restricted to circumstances where (1) firms are of a fly-bynight kind, (2) firms are playing end games, and (3) intergenerational learning is
negligible.”
Employees with a “fixed wage” arrangement are protected from the month-tomonth, or year35
to-year fluctuations in the rents generated by their specific skills. But there are few if
any protections for employees if the present value of the stream of rents their specific
skills can generate falls below the present value of the stream of wage premia they are
receiving. 24 they face. This is because it is much harder to enforce “fairness” in an
employment agreement whose terms can be renegotiated as business conditions facing
the firm vary. As Milgrom and Roberts note: “The firm's management may be
tempted to exaggerate financial difficulties in order to justify paying lower wages to
workers.”
36 On the other hand, totally insulating employees against risks might discourage
them from doing the things that are under their control to pull resources out of lowervalue investments and move them to higher-value investments, by retraining, for
example. B. Monitoring. The disincentive effects of mechanisms that shelter
employees from some of the risks inherent in making specialized investments can
possibly be counteracted with more intensive monitoring. But monitoring, in turn,
can present serious measurement, verification, and evaluation problems. The monitor
must focus on the measurable dimensions of performance, 37 which may lead
employees to focus on those “monitored” dimensions to the exclusion of
nonmeasurable dimensions that may also be important to productivity.
38
See Milgrom and Roberts (1992), p. 334. See also Rosen (1985), for a good
summary of risk36
sharing issues in implicit employment contracts.
Hashimoto (1981) argues that investments in firm-specific human capital are
likely to present
37
such a severe problem in evaluating, verifying and agreeing on workers'
productiveness that the monitoring solution cannot be effective by itself. He models
Becker's argument that employee and employer should share the costs and returns
from such investments, using a sharing rule that explicitly gives employees a
fluctuating share in the rents, rather than a fixed wage.
See Holmstrom and Milgrom (1991). These authors focus on the problems
caused by pay for
38
performance schemes that tie compensation to quantifiable measures of performance
but exclude consideration of other, less quantifiable indicators of performance. They
argue that monitoring can help to counteract the potential for perverse incentive
effects from such schemes. But monitoring is subject to the same kinds of perverse
incentive problems as pay for performance schemes if all of the dimensions of
performance are not equally observable, quantifiable, and verifiable, and therefore
susceptible to monitoring.
25
C. Quantifying the value of residual claims. One of the contracting problems raised
by
investments in human capital in general, and by specific human capital in particular, is
the problem
of understanding and quantifying all the forms that the returns to those investments
can take.
Whereas the returns to physical capital investments can generally be measured in
monetary terms,
some of the returns to investments in human capital may take other forms. Human
capital may not
depreciate with use, for example, but may, instead, appreciate. Like muscles,
knowledge and skills
that are used may build on themselves and become more useable. If so, the returns to
tenure may go
well beyond just the returns from skills accumulated up to a particular point in time
(and by
extension, the losses from premature job separations may be much larger than that
implied by the
stream of rents being generated today). The losses may also include a component that
is like an
option. If the employee stays with his present employer, he will have an opportunity
to acquire
skills tomorrow that build on the skills acquired through today. Those skills would
generate an
additional stream of returns on top of the stream of returns from the skills
accumulated through
today. If the employment relationship is prematurely severed, that option value is lost.
The complex nature of the returns to human capital may make it impossible, Milgrom
and
Roberts note, “to identify any individual or group that is the unique residual claimant
[in a firm], or,
indeed, to identify the benefits and costs accruing to any decision and so compute the
residuals.”
39
The difficulty of computing and assigning residuals complicates the problem of
bargaining between
any employee and the firm over the allocation of residual claims, or over any scheme
of payments
See Milgrom and Roberts (1992), p. 315. See also Holmstrom (1982) for a
general discussion
39
of a related problem with team production, the problem of allocating returns from
team efforts in ways that discourage “free-riding” by individual members of the team.
26
that might be devised to encourage both parties to the relationship to take into account
the impact of
decisions by one on the other.
D. A Counter Argument. Dow argues that employees don't generally make any
sacrifices
when they acquire firm-specific skills -- that, instead, they acquire such skills as a
byproduct of
participating in production tasks, a job they are being paid to do anyway. Thus, he
says, “there is no
need for explicit contracts whereby workers commit themselves to invest in
specialized skills,” and
therefore, the “non-contractibility of investments in human capital does not raise an
insuperable
obstacle for capitalist firms.” But this argument seems too facile. Even if employees
don't have to
40
make great sacrifices to acquire firm-specific skills, those skills may still have value,
to the firm, or
to the employee or to both, and hence, it may be socially valuable to provide
protective contractual
arrangements or governance devices.
41
See Dow, “Why Capital Hires Labor: Reply,” undated working paper, University
of Alberta.
40
If someone inherited a piece of land, the fact that they acquired that piece of land
without
41
expending any special effort or resources does not mean that the value of the land
should not be protected.
27
IV. Institutional Arrangements that Address These Contracting Problems. Since the
contracting problems surrounding investments in firm-specific human capital are so
large and
pervasive, it should not be surprising to find that the providers of human and financial
capital have
found other, non-contractual mechanisms for encouraging and protecting firm-specific
investments.
Some of these mechanisms have been studied by labor economists but while versions
of them are
found in most large corporations, they have not generally been treated by economists
as being part of
the institutional nature of the firm itself.
A. Long-term employment relationships. Customs and practices that encourage
longterm employment relationships have a variety of benefits that support, or are
perhaps complementary
with investments in firm-specific human capital. Milgrom and Roberts cite “an
increased
opportunity to invest profitably in firm-specific human capital, the greater efficacy of
efficiency
wage incentives contracts in long-term relationships, and the enhanced ability to make
an accurate
assessment of an employee's contributions to long-term objectives by monitoring
performance over a
longer period of time.” To this could be added reputational considerations. As
argued above, good
42
reputations are more valuable the longer the time-horizons of the contracting parties,
so that in a
longer-term relationship, both sides to any given transaction within that relationship
will have
stronger incentives to perform fully.
B. “Hostages” or Performance Bonds. A “hostage” is anything of value that is
pledged
by one party to a transaction, and that will be forfeited to the other party if the first
party fails to
perform his or her end of the contract. One version of the “efficiency wage”
arguments, for
example, is based on a “hostage” argument. That argument says that workers accept
wages that are
See Milgrom and Roberts (1992), p. 363. The classic discussion of the role
played by long42
term employment relationships is Doeringer and Piore (1971).
28
lower than their opportunity cost in the early years of their employment relationship,
and that this
serves as a performance bond that is later repaid in the form of wages that are higher
than their
opportunity cost.
Severance pay commitments, and their gilt-edged cousins, golden parachutes, can be
thought
of as “hostages” provided by the employer. Penalties for certain kinds of changes in
the contract
terms may perform a similar function. Milgrom and Roberts argue, for example, that
employment
contracts might be designed to impose a penalty of some sort on the employer for
invoking a claim
of hard-times in an effort to negotiate lower wages, so that the employer will not be
tempted to use
this claim frivolously in negotiations.
43
The administration and enforcement of performance bond agreements, however,
requires
that third parties be able to observe and verify certain measures of performance, and
certain
triggering events. Hence performance bonds by themselves seem like poor candidates
for solving the
contracting problems presented by the accumulation of specialized human capital
unless they are
imbedded in institutional arrangements that also foster trust, or that make reputations
valuable.
C. Job ladders, career paths, and seniority rules. A number of scholars have argued
that career paths and job ladders are important mechanisms for encouraging
employees to make
investments in firm-specific human capital and for ensuring that the firm shares the
rents generated
by those investments with employees. Seniority rules are a related mechanism that
provides some
44
protection for employees from the possibility that the firm will renege on its implicit
agreement to
compensate the employee for his or her firm-specific investments by paying them a
higher wage
See Milgrom and Roberts (1992), p. 334.
43
See, for example, Koike (1990), and Prendergast (1993).
44
29
during their later years, perhaps even a wage that exceeds their productivity during
those years.
Seniority rules protect the high-tenured worker by requiring the firm to lay off lowtenured workers
first.
Both seniority rules and job ladders can be seen as mechanisms that help ensure that
the
employee will be appropriately compensated for his or her investments over time. But
such
promises would have no incentive benefit if employees did not believe that the
employing
organization would continue to exist over the relevant period of time. So these
mechanisms are not
useful by themselves, but must be embedded in a relationship that is understood to be
long-term,
with an entity that is long-lived.
D. Unionization. One effect of unions is to substitute “voice” for the right of “exit”
(i.e.,
the “at will” contract) for both sides to the employment relationship by providing
protections for
employees such as protection from dismissal, except “for cause.” Other terms in the
typical
collective bargaining agreement, in turn, help prevent firms from driving out an
unwanted employee
without actually dismissing him. These take the form of rules designed to protect
wages, benefits,
and job assignments, as well as protection against layoffs. These sorts of protections
have been
criticized because they impose rigidities that can have negative implications for
efficient adaptation
to changed circumstances. But these costs must be weighed against incentive
benefits union
45
agreements may provide.
See Epstein (1985), p. 147. “The organizational difficulties of collective
bargaining are the
45
price that union workers pay in order to extract monopoly profits, or at least firmspecific rents, from their employers and through them the consuming public at large,”
Epstein claims. See also Klein, Crawford and Alchian (1978). Epstein's argument
involves an unstated assumption that the “firmspecific rents” at stake belong to the
employers rather than to the employees in the first place.
30
E. Corporate culture, norms, and goals. Milgrom and Roberts note that “workable
principles and routines . . . create shared expectations for group members.” The
advantages of such
principles are that they “help guide managers in making decisions”; that they provide
“a set of clear
expectations for everyone in the organization”; and that they “provide a set of
principles and
procedures for judging right behavior and resolving inevitable disputes.” As Kreps
(1990) first
46
argued, these aspects of corporate culture may serve as “focal points” around which
participants in
the firm can arrive at stable patterns of interacting that are Pareto superior to patterns
they might
lapse into without the benefit of the common norms. Hence, corporate culture can
help to support
47
investments in firm-specific human capital by fostering trust.
Corporate culture can also be seen as part of the firm-specific capital of the firm -- the
organizational capital if you will. Nelson and Winter argue that the knowledge of
how to do things
is often implicit in the routines that make up the daily activities of the people in the
firm. As such, 48
it is neither articulable, nor alienable, but is embodied in the people and in their
relationships to each
other. Similarly, Jacoby writes about the “socialization of the workplace itself, which
relies on
consensual methods of inculcating norms and goals, such as ideologies or authority
that must be
seen as legitimate if they are to be persuasive.”
49
F. Ownership and control rights.
Milgrom and Roberts (1992), p. 265.
46
Kreps (1990) provides an extensive analysis of the role of corporate culture using
game47
theoretic arguments.
Nelson and Winter (1982).
48
See Jacoby (1990), p. 332.
49
31
Hart has argued that “ownership” should be defined as possession of “residual”
control
rights -- the right to make all decisions that have not been specified by contract.
Organizing a
50
transaction within a single firm has the effect of assigning all residual control rights
over the
nonhuman assets used in that transaction to a single entity, Hart suggests. And it is
this
51
concentration of control rights that makes it possible to accomplish things that could
not be
accomplished through market-mediated transactions among individual actors, each of
whom
controlled only the physical assets with which he or she worked. In particular, Hart
has noted, “ex
post residual rights of control will be important because, through their influence on
asset usage, they
will affect ex post bargaining power and the division of ex post surplus in a
relationship. This in
turn will affect the incentives of actors to invest in that relationship.”
52
Hart has argued that “co-specialized” assets should be owned in common. If they are
not,
then the separate parties who own each asset will have reason to fear that the other
parties will
expropriate “too much” of the rents earned by the assets, and will tend to underinvest.
But of
53
See Hart (1989).
50
Wiggins asserts that “in principle, parties could agree to any contract between
firms that could
51
be entered into internally.” See Wiggins (1991), p. 661. This is the inverse of the
proposition put forth by Coase that, in principle, all of the arrangements that are
available to contracting parties outside the firm should be available within the firm,
and then some. By extension, organizing economic activity within firms would
always be more efficient than organizing activity between autonomous units. Coase
asked what it is that makes organizing economic activity within firms more efficient
in some circumstances, and organizing activities through contractual or market
agreements between autonomous parties more efficient in other circumstances. See
Coase (1937).
Wiggins suggests that the difference has to do with which form makes it easier or less
costly to enforce the contracts or terms of the relationship. Almost by definition,
“residual” control rights are the non-contractible ones. Under common “ownership,”
these residual control rights reside with the same party.
See Hart (1989), p. 1766.
52
See Hart (1989), p. 1757-1774.
53
32
course, neither the firm, nor any other participant in the enterprise that the firm
directs, can “own”
the human capital that may be co-specialized with the other assets of the firm. Hence,
where firmspecific human capital is important, such arguments might in some cases
point toward employee
control of the enterprise, or at least participation in management, rather than capitalist
ownership
and control.
54
1. Labor-managed firms. A number of scholars were inspired by the Yugoslavian
experiment with labor-managed firms in the 1960s to consider the advantages and
disadvantages of
organizing production in this way. This produced a small, but lively debate in which
neoclassical
economists argued that employee-controlled firms would be inefficient for a variety of
reasons.
55
Various scholars have answered these criticisms by pointing out that, in each case, the
supposed
inefficiencies are a product of peculiar modeling assumptions made by the critics.
But, as an
56
empirical matter, employee-controlled firms remain rare, and absent obvious legal
restrictions
In noting the advantages of the partnership form of organization, Milgrom and
Roberts point
54
out that “human capital is not easily tradable, and if the residual returns on that capital
belong to the humans who embody it, then the usual arguments about ownership
rights suggests that the residual control should be assigned to them too.” See
Milgrom and Roberts (1992), p. 523. See also Hansmann (1988), p. 292, Putterman
and Kroszner (1995), p. 19, and Blair (1995), for discussions of employee ownership
as a mechanism for protecting investments in firm-specific human capital.
Vanek (1977) and Meade (1972) argued that labor-managed firms would
maximize net
55
revenues per worker rather than maximizing profits, for example. Jensen and
Meckling (1979), Furobotn and Pejovich (1974), and Vanek (1977) argued that
employee-controlled firms would not have the right incentives to adequately maintain
their physical capital. Williamson (1975, 1985) argued that democratically-run firms
would be inefficient because hierarchies are needed for efficient processing of large
amounts of information.
See e.g., Putterman (1984), Wolfstetter, Brown and Meran (1984), Ellerman
(1986), and Dow
56
(1993).
33
against them, economists generally assume take their absence to mean that this form is
not
economically viable for a variety of reasons.
57
2. Labor participation in management. Corporate governance systems in Japan
and Europe seem to feature institutional arrangements that provide mechanisms by
which employees
are given a direct voice in management. Although it is beyond the scope of this paper
to discuss
these alternative systems, Japanese scholars, especially, have credited these
arrangements with
providing incentives and protection for employee investments in firm-specific human
capital.
58
Germany's co-determination system has also attracted the attention of policy-makers
and scholars in
this regard.
59
3. Equity ownership by employees.
Compensating employees with equity stakes in corporations might provide a
mechanism for
fostering and protecting investments in firm-specific human capital. Equity
ownership by employees
serves as a kind of “hostage,” helping to make the firm's promise to share in the rents
credible. It
also gives employees some control rights (by virtue of their equity holdings rather
than by virtue of
their status as employees), while at the same time, helping to align their interests with
those of
outside equity holders. And if equity claims are substituted for the wage premium that
firm-specific
human capital supposedly earns, the wages will come closer to reflecting opportunity
cost, and
thereby send the correct economic signals to decision-makers within the firm to guide
hiring and
firing decisions.
See Dow and Putterman (1997) in this collection.
57
The literature on this is large, but see, e.g., Aoki (1988), and Matsuyama (1996).
58
See, e.g., Baums and Frick (1997) and Pistor (1997) in this collection.
59
34
There is substantial evidence that the use of equity-based compensation systems is
growing
in U.S. corporations, although no definitive empirical studies have linked employee
ownership in
publicly-traded firms to investments in firm-specific human capital.
60
4. A Case Against Employee Ownership. Hansmann has argued that the
disadvantages of collective decision-making by heterogeneous employees, however,
might easily
outweigh the advantages of common ownership of the capital and labor inputs. He
argues that the
feature that distinguishes employee-owned enterprises from those owned by outside
investors is the
homogeneity of the workforce, rather than the specificity of the human capital
possessed by that
workforce. Other scholars have argued that the fact that capitalists have more wealth
and better
61
access to credit markets than workers do, and the fact that capitalists can diversify
risks better than
workers can, also argue against employee ownership.
See Blasi and Kruse (1991) for the most comprehensive evidence of the growth of
employee60
ownership in publicly-traded firms.
See Hansmann (1995).
61
35
V. New Thinking about the Theory of the Firm.
The foregoing review of the role played, and problems raised by, investments in firmspecific
human capital suggests that the nature of the employment relationship is, indeed,
central to
understanding the nature of the institutional arrangements that are at the essence of
modern, large
corporations. The idea that the firm is a “nexus of contracts,” was a significant insight
that helped
push our understanding of firms forward because it got scholars thinking about the
terms of the
relationships among the various participants in firms. But scholarly work on this idea
has perhaps
been hung up for too long in a fixation on one particular relationship (that between
shareholders and
managers) and on an approach to modeling that relationship (principal-agent models)
that is too onesided.
Economic theorists, however, are moving in some new directions that acknowledge
more
directly the complex nature of the way employees participate in firms. These
theoretical
contributions take steps toward defining the firms themselves in terms of the
institutional
arrangements developed to elicit contributions by employees to the joint productive
effort of the
enterprise. If the full range of contributions needed could be adequately elicited
through market
relationships or explicit contracts, perhaps they would be. But it is the very fact that
they can't that
calls forth complex organizational forms such as modern corporations.
A. The Firm as a System of Incentives. Holmstrom and Milgrom (1994) propose that
firms should be viewed as incentive systems. Their model explores why “the
attributes of an
employment relationship differ in so many ways from the attributes of a contractor
relationship.” In
comparing the terms on which in-house sales agents typically operate, to the terms on
which
independent sales agents typically operate, for example, Holmstrom and Milgrom note
that, relative
to relationships with subcontractors, employment relationships typically involve
lower-powered
36
incentives (a fixed base salary and lower commissions, for example), ownership of
key assets by the
employer (rather than by the employee), and more restrictions on mode of operation of
the
employee.
Holmstrom and Milgrom's contribution utilizes a multitask principal-agent model
designed
to address the problems that arise when the tasks the worker is supposed to do are
multidimensional,
and performance difficult to measure in some or all of those dimensions. As the
authors had
discussed in previous work, when agents must perform a number of tasks, and their
choices about
effort and allocation of their time can affect many dimensions of the firm's
performance, highpowered incentive structures that reward performance in some
dimensions (but neglect performance
in other dimensions) can greatly distort the behavior of the agent.
62
Of course, a key feature that distinguishes agents who are “in” the firm from agents
who are
on the outside and merely contracting with the firm, is the structure of the
compensation agreement.
In particular, compensation for contractors generally provides for task-specific
payments, with all
risks of nonperformance borne by the agent, whereas with employees, such risks are
generally
pooled (and borne collectively by the firm itself, and ultimately by various participants
in the firm),
so that the agent is paid a regular wage or salary for the duration of employment,
regardless of the
actual tasks performed.
The Holmstrom and Milgrom model implies that, under certain conditions, an optimal
incentive structure “may require the elimination or muting of incentives which in a
market
relationship would be too strong.” Thus, they conclude, “the use of low-powered
incentives within
See Holmstrom and Milgrom (1991).
62
37
the firm, although sometimes lamented as one of the major disadvantages of internal
organization, is
also an important vehicle for inspiring cooperation and coordination.”
63
Within the world of insurance agents that Holmstrom and Milgrom consider, they find
that
the choice between structuring the relationship as an employment one, versus
structuring it as one of
an independent contractor type, appears to be driven by the relative ease or difficulty
of measuring
key aspects of performance, more than by the extent of investments in firm-specific
human capital.
But it seems unlikely that this factor drives this choice in other cases. Consider
production
line workers, for example. Assembly line workers who work on large, highly capitalintensive
automated assembly lines are typically paid hourly wages, and a variety of other
institutional
arrangements are used (such as pension funds and collective bargaining) to discourage
turnover. By
contrast, workers in garment factories are more likely to be paid piece rates, turnover
rates in sweat
shops are high, and there are fewer institutional arrangements designed to reduce
turnover. In other
words, apparel workers are often compensated and treated more like subcontractors
than employees.
In both cases, the activities of the worker should be very easy to measure. But in the
garment
factory, where apparel is assembled at separate sewing machines, individual workers
can set their
own pace, whereas in large automated factories, individual workers cannot set their
own pace, but
must learn to function at a pace set for them by the machines and by the rest of the
team with which
they work. Holmstrom and Milgrom's model might be used to test the hypothesis that
the differences
in compensation systems and institutional arrangements between, say, auto factory
workers and
garment factory workers, are accounted for by the fact that workers on automated
assembly lines
See Holmstrom and Milgrom (1994), p. 989.
63
38
must make a higher level of investment (that is, exert more “effort”) in learning to
work with the
particular equipment in the factory and with the particular teammates on the assembly
line.
The Holmstrom and Milgrom paper illuminates the importance of considering the
whole
mix of incentives facing employees, and moves theory away from viewing firms as
“bundles of
assets,” toward a view of firms as constellations of institutional arrangements
designed to provide
appropriate incentives where cooperation and coordination are especially important.
Their modeling approach, however, falls within the principal-agent paradigm, and as
is
generally the case with principal-agent models, it does not take into account the
incentives facing the
principal to renege on the promised payment scheme, or to alter the job design in ways
that reduce
the payoff to the agent ex post. It also does not explain two other features that
distinguish the
employment relationship from the independent contractor relationship -- features that
have been
cited as evidence that investments in firm-specific human capital are important.
These are the
longevities typically observed in the employment relationship relative to independent
contracting
relationships, and the wage premia associated with tenure.
B. The Firm as a Nexus of Specific Investments
Another very new contribution to the literature, by Rajan and Zingales (1996), treats
firmspecific investments, especially in human capital, as the defining feature of a firm.
This approach
64
is reminiscent of Aoki (1984), who defined the firm as “an enduring combination of
firm-specific
The authors suggest that a firm should be defined not as a “nexus of contracts,” as
previous
64
scholars have proposed, but as a “nexus of specific investments.” (Although the
authors have produced a revised version of the working paper cited in this article, with
the same title, but dated November, 1996, the revised version compresses some
important arguments that I want to emphasize in this article. The authors have
explained to me that they intend to pursue those compressed ideas in other papers.)
39
resources.” Rajan and Zingales use an optimal-contract modeling approach that is
similar to, and
65
builds on the approach used by Grossman and Hart in their landmark piece (discussed
above).
66
In Rajan and Zingales's model, there is a physical asset that is specific to the
enterprise, and
two individuals. The total productivity of the enterprise will be maximized if both
individuals make
specific investments in human capital. But each individual must have access to the
physical asset in
order to “specialize.” If either individual fails to specialize, an unspecialized outsider
can be
substituted for that individual without loss of total productivity.
Rajan and Zingales distinguish between “ownership” and “power.” “Ownership” of
the
enterprise, in their model, gives the owner the right to exclude other individuals from
access to the
physical asset, and the right to sell the physical asset to some third party. These rights
give the
67
“owner” significant “power” in bargaining over the ex post distribution of rents.
But participants can also get “power” in another way. Investment by either individual
in
firm-specific human capital also gives that individual bargaining power in the
relationship, because
his investment in human capital means that there will be more total rents to share if he
stays in the
coalition and uses his human capital in the enterprise.
Aoki (1984), p. 119. Aoki argued that firms should be regarded as combinations
of specific
65
labor and capital, and that management should be viewed as mediating between these
two interests in making decisions about output levels, investments, and sharing of
firm-level rents. His model takes the relative power of labor and capital as
parameters, and models the decision-making process of management. Rajan and
Zingales, as we shall see, derive a similar structure for firms from more primitive
assumptions about the nature of the technology and the incentives of the individual
participants.
Grossman and Hart (1986).
66
Grossman and Hart (1986) define ownership only in terms of the ability of the
owner to
67
exclude others from access. Their model does not allow for the owner to sell the asset
into any kind of external market.
40
Grossman and Hart (1986), and Hart and Moore (1990) argued that ownership of the
physical asset would increase the ex ante incentive for the owner to make the optimal
investments in
human capital. But Rajan and Zingales point out that ownership of the physical asset
also enables
the owner to sell the asset, or to share in the rents from the enterprise even if he fails
to make firmspecific investments. Hence ownership rights over the physical asset, in
their model, has a doublededged effect. It increases the bargaining power the owner
has, and therefore increases his incentive
to “specialize” by assuring him that his share of the rents generated by the enterprise
will not be
expropriated ex post. This is the Grossman-Hart, and Hart-Moore effect. But it also
raises his
opportunity cost for specializing, since he can extract rents even if he doesn't
specialize.
If the negative effects of ownership by either individual dominate the positive
incentive
effects, Rajan and Zingales show that the optimal investment decisions and production
levels cannot
be achieved if either of the two potential “specializers” own the physical asset. But,
remarkably, if
the physical asset is owned by an otherwise passive third party, optimal investment
decisions and
production levels can still be achieved. In this situation, the two individuals who want
to participate
in the firm would form a coalition and bid collectively for access to the asset, and the
right to use the
asset in production. Third-party control over the physical asset helps to encourage
both individuals
to make the optimal firm-specific investments, because it, in effect, enables the two
individuals to
make a binding commitment not to use control over the asset strategically to try to
extract rents from
the other individual.
Rajan and Zingales have, thus, brought us full circle, back to the Alchian and Demsetz
story
about the importance of team production, and the need for a third party to monitor the
inputs of the
team members. But because they assume that the individual members of the team are
not generic
inputs, but specialists who got that way because they invested in learning things that
only have value
41
when used by this particular team, they reach a very different conclusion about the
division of rents
from team production, as well as about the role played by the third party “monitor.”
Rajan and Zingales's third party has no special knowledge or insights about how the
work is
divided up between the two individuals in the coalition, nor how they divide up the
rents between
them. Their third party is assumed to get an arbitrarily small fraction of the total rents,
and his only
task is simply to select from among multiple coalitions bidding for access rights to the
physical
assets. He naturally selects the coalition that will produce the highest total rents, the
bulk of which
go not to the third party monitor (as in Alchian and Demsetz), but to the coalition
members who
have invested in specialized human capital.
“Before investment [in specialized human capital] takes place, the firm is defined by
who
holds the ownership rights to the physical assets that are required for production and
by who is given
access to the physical assets,” Rajan and Zingales argue. “After specific investment
has been
undertaken, the firm is defined by the ownership of the physical assets and the power
that accrues to
those who have made specific investments.”
68
Rajan and Zingales interpret their model as offering an economic rationale for the
separation
of “ownership” from “control,” although this seems like a curious interpretation since
they define
“ownership” to mean control rights over the use of physical assets. One right that
outside
shareholders of large corporations definitively do not have is control over the physical
assets of the
corporation.
An alternative interpretation of Rajan and Zingales's work might be that it provides
insight
into the role played by independent boards of directors with fiduciary obligations to
their firms. The
The authors go on to argue that, with their model, there can be a firm even if there
are no
68
physical assets, although in such a case there is also no role for a third party “owner.”
42
function of the “passive” third-party “owner” in their model is not to “maximize the
value” of his
stake. In fact, the third party owner in this story is restricted to receiving an arbitrarily
small return
because he does not provide anything critical to production. His role could be played
by anyone
(except, notably, any of the active participants in the enterprise). His function instead
is to keep
control of the assets out of the hands of any of the active participants in the firm,
precisely so that
those active parties won't use control over the assets to gain strategic advantage for
themselves
at the expense of the other participants, and thereby cause the coalition to fall apart.
C. Future Directions.
Rajan and Zingales have taken a significant step toward integrating models of the
employment relationship (and the associated incentive issues raised by investments in
firm-specific
human capital) into a theory of the firm. But their model is still limited by the fact
that it follows the
two-period structure of most bargaining models. In such models, contracts are
written, investment
69
decisions are made, production proceeds, and rents are realized and divided up. End
of story. There
is no second round, let alone third or fourth or more rounds, so that there is no place
in the model for
reputations to be built up, or for learning from experience, or for investments made in
previous
rounds to expand the options for the participants in subsequent rounds.
Models that have such features can become intractable very quickly, and theorists who
have
struggled with repeated game models find that they are plagued by multiple equilibria,
and are often
very sensitive to assumptions about who has what information when. Nonetheless,
the basic
70
insight from infinitely repeated game models is that the chance to benefit from a
relationship in the
Technically, there are three periods in the Rajan and Zingales model, but that is
because they
69
have made the decisions by the two individuals to specialize sequential instead of
simultaneous.
See Kreps (1996) for an interesting discussion of these issues.
70
43
future can mitigate tendencies that parties to the relationship might have in the present
to attempt to
expropriate short run returns.
From the perspective of repeated games, each act of self-restraint on the part of
participants
in the firm can be seen as a “firm-specific investment” whose value can be realized if
the coalition
stays together, but not if it falls apart. The cumulative result of a large number of such
acts of selfrestraint could represent a sizeable investment in a type of firm-specific
human capital that we might
call “trust,” or “culture.” As in Rajan and Zingales, the firm can be viewed as a nexus
of these
investments, and for the full value of the investments to be realized, the coalition of
key participants
in the firm must be kept together. Also, as in Rajan and Zingales, keeping the
coalition together
may require the services of an otherwise neutral third party with critical control rights.
Again, the
role is more suggestive of an independent board of directors than of outside
shareholders.
VI. Implications for Legal Scholarship.
In the last two decades, the idea that a corporation is a “nexus of contracts” has
strongly
influenced legal scholarship. The idea was seen as a substitute for an “entity” view of
the
corporation, a view which had been prevalent in legal theory prior to about the 1970s.
Under the
entity view of the firm, a corporation is understood to be something apart from each of
its
participants that managers could raid or abuse if they were not subject to constraints
imposed by
their fiduciary duties and corporation law.
The “nexus of contracts” view, by contrast, offered a view of corporations that
stresses the
willing participation by all parties, and the ability of all parties to protect their
interests by
contractual means. The legal implication of the nexus of contracts view was taken to
be that if the
courts and the legislatures would restrict themselves merely to enforcing contracts, the
interplay of
freely contracting individuals and free markets would lead individuals to enter into
relationships that
44
optimally balanced claims to returns and control rights. Contracts freely entered-into,
moreover,
could be assumed to be economically efficient.
71
In this stark form, however, the contractarian view fails to recognize that, no matter
how
beneficial they might be, certain kinds of multilateral, and multidimensional
relationships and
agreements among individuals may only be possible in a legal environment that grants
separate legal
status to the entity that serves as the repository of the specific investments involved in
the
relationship. It may also be necessary that the law assign fiduciary responsibilities to
the individuals
whose job is to govern this entity.
The theoretical work discussed above suggests that, as we come to understand better
the full
and complex dimensions of the contracting problem involved in organizing
production, we may
develop a renewed appreciation of the entity view of firms in the law, and with that,
perhaps even of
the role of legal constraints on managerial behavior and of enforceable fiduciary
obligations on the
part of directors.
See, Butler (1989) for a discussion of the “nexus of contracts” view of the firm as
the basis for
71
a contractarian approach to corporate law.
45
Bibliography
Alchian, Armen A., and Harold Demsetz. 1972. “Production, Information Costs, and
Economic Organization.” American Economic Review 62: 777-795.
Aoki, Masahiko. 1984. The Co-operative Game Theory of the Firm, Oxford:
Clarendon Press.
------------. 1988. Information, Incentives, and Bargaining in the Japanese Economy,
Cambridge: Cambridge University Press.
Arrow, Kenneth J. 1985. “The Economics of Agency.” In Principals and Agents: The
Structure of Business, edited by John W. Pratt and Richard Zeckhauser, 37-51.
Boston: Harvard Business School Press.
Baker, George, Robert Gibbons, and Kevin Murphy. 1996. “Implicit Contracts and the
Theory of the Firm.” Working Paper, April, 1996.
Baums, Theodor, and Bernd Frick. 1997. “Co-determination in Germany: The Impact
of Court Decisions on Stock Price,” (in this collection)
Becker, Gary S. 1964. Human Capital: A Theoretical and Empirical Analysis, With
Special Reference to Education. New York: National Bureau of Economic Research.
Blair, Margaret M. 1995. Ownership and Control: Rethinking Corporate Governance
for the Twenty-First Century. Brookings.
_______________. 1996. Wealth Creation and Wealth Sharing: A Colloquium on
Corporate Governance and Investments in Human Capital. Brookings.
Blasi, Joseph Raphael, and Douglas Lynn Kruse. 1991. The New Owners: The Mass
Emergence of Employee Ownership in Public Companies and What it Means to
American Business, HarperCollins Publishers.
Butler, Henry N. 1989. “The Contractual Theory of the Corporation,” George Mason
U.
L.Rev.11(Summer): 99.
Castanias, Richard P., and Constance E. Helfat. 1992. “Managerial and Windfall
Rents in the Market for Corporate Control.” Journal of Economic Behavior and
Organization 18 (July): 153184.
Coase, R. H. 1937. “The Nature of the Firm.” Republished in Oliver Williamson and
Sidney Winter, eds., The Nature of the Firm, Cambridge: Cambridge University Press,
1991.
Demsetz, Harold. 1991. “The Theory of the Firm Revisited,” in Oliver Williamson
and Sidney Winter, eds., The Nature of the Firm, Cambridge: Cambridge University
Press.
46
Dickens, William T., and Kevin Lang. 1993. “Labor Market Segmentation Theory:
Reconsidering the Evidence.” In Labor Economics: Problems in Analyzing Labor
Markets, edited by William Darity, Jr., 141-180. Boston: Kluwer Academic
Publishers.
Doeringer, Peter B., and Michael J. Piore. 1971. Internal Labor Markets and
Manpower Analysis.
Lexington, Massachusetts: D.C. Heath and Company.
Domar, Evsey. 1966. “The Soviet Collective Farm as a Producer Cooperative,”
American Economic Review, Vol. 56, pp. 734-57.
Dow, Gregory K. 1993. “Why Capital Hires Labor: A Bargaining Perspective.”
American Economic Review 83 (March): 118-34.
Dow, Gregory K., and Louis Putterman. 1997. “Why Capital (Usually) Hires Labor:
An Assessment of Proposed Explanations,” (in this collection).
Ellerman, David P. 1986. “Horizon Problems and Property Rights in Labor-Managed
Firms,” Journal of Comparative Economics, Vol. 10, pp. 62-78.
Epstein, Richard A. 1985. “Agency Costs, Employment Contracts, and Labor
Unions.” In Principals and Agents: The Structure of Business, by John W. Pratt and
Richard Zeckhauser, 127148. Boston: Harvard Business School Press.
Fama, Eugene F., and Michael C. Jensen. 1983. “Separation of Ownership and
Control.” Journal of Law and Economics 26 (June): 301-325.
Furubotn, Erik G., and Svetozar Pejovich. 1974. “Property Rights and the Behavior of
the Firm in a Socialist State: The Example of Yugoslavia.” In The Economics of
Property Rights, edited by Furubotn and Pejovich. Cambridge, Mass.: Ballinger
Publishing Co.
Grossman, Sanford J., and Oliver D. Hart. 1986. “The Costs and Benefits of
Ownership: A Theory of Vertical and Lateral Integration.” Journal of Political
Economy 94 (August): 691-719.
Hansmann, Henry. 1988. “Ownership of the Firm,” Journal of Law, Economics, and
Organization, Vol. 4, No. 2, Fall, pp. 267-305.
Hansmann, Henry. 1996 (forthcoming). The Ownership of Enterprise. Harvard
University Press.
Hart, Oliver. 1989. “An Economist’s Perspective on the Theory of the Firm.”
Columbia Law Review 89: 1757-74.
Hart, Oliver, and John Moore. 1990. “Property Rights and the Nature of the Firm,”
Journal of Political Economy 98 (6): 1119-58.
47
Hashimoto, Masanori. 1981. “Firm-Specific Human Capital as a Shared Investment.”
American Economic Review 71 (June): 475-482
Holmstrom, Bengt. 1982. “Moral Hazard in Teams.” Bell Journal of Economics, Vol.
13, No. 2, Autumn, pp. 324-340.
Holmstrom, Bengt, and Paul Milgrom. 1991. “Multi-task Principal-Agent Analyses:
Incentive Contracts, Asset Ownership, and Job Design.” Journal of Economics and
Organization 7 (Special Issue): 24-52.
________. 1994. “The Firm as an Incentive System.” American Economic Review 84
(September): 972-91.
Jacobson, Louis S., Robert J. LaLonde, and Daniel G. Sullivan. 1993. “Earnings
Losses of Displaced Workers.” American Economic Review 83 (September): 685709.
Jacoby, Sanford M. 1990. “The New Institutionalism: What Can It Learn from the
Old?” Industrial Relations 29 (Spring): 316-359.
Jensen, Michael C., and William H. Meckling. 1976. “Theory of the Firm: Managerial
Behavior, Agency Costs and Ownership Structure.” Journal of Financial Economics 3
(October): 305-60.
________. 1979. “Rights and Production Functions: An Application to LaborManaged Firms and Codetermination.” Journal of Business 52 (October): 469-506.
Katz, Lawrence F., and Lawrence H. Summers. 1989. “Industry Rents: Evidence and
Implications.” Brookings Papers on Economic Activity: Microeconomics: 209-275.
Klein, Benjamin, Robert A. Crawford, and Armen A. Alchian. 1978. “Vertical
Integration, Appropriable Rents, and the Competitive Contracting Process.” Journal
of Law and Economics
21 (October): 297-326.
Koike, Kazuo. 1990. “Intellectual Skill and the Role of Employees as Constituent
Members of Large Firms in Contemporary Japan.” in The Firm as a Nexus of Treaties,
edited by Masahiko Aoki, Bo Gustafsson, and Oliver E. Williamson, 185-208. Sage
Publications.
Kreps, David M. 1990. “Corporate Culture and Economic Theory.” In Perspectives on
Positive Political Economy, edited by James E. Alt and Kenneth A. Shepsle, 90-143.
Cambridge University Press.
_____________. 1996. “Markets and Hierarchies and (Mathematical) Economic
Theory.” Working Paper.
Krueger, Alan B., and Lawrence H. Summers. 1988. “Efficiency Wages and InterIndustry Wages Structure.” Econometrica 56 (March): 259-293.
48
Masten, Scott E., James W. Meehan, and Edward A. Snyder. 1989. “Vertical
Integration in the U.S.
Auto Industry: A Note on the Influence of Transaction Specific Assets.” Journal of
Economic Behavior and Organization 12 (October): 265-273.
Matsuyama, Kenji, 1996. “Corporate Governance: A Perspective on the Japanese
System,” Japan Research Institute, unpublished paper. October.
Meade, J. E. 1972. “The Theory of Labor-Managed Firms and of Profit Sharing.”
Economic Journal 82 (March supplement): 402-28.
Milgrom, Paul, and John Roberts. 1992. Economics, Organization and Management.
New Jersey:
Prentice Hall, Inc.
Monteverde, Kirk, and David J. Teece. 1982. “Supplier Switching Costs and Vertical
Integration in the Automobile Industry.” The Bell Journal of Economics 13 (Spring):
206-213.
________. 1982. “Appropriable Rents and Quasi-Vertical Integration.” The Journal
of Law and Economics 25 (October): 321-328.
Nelson, Richard R., and Sidney G. Winter. 1982. An Evolutionary Theory of
Economic Change.
Harvard University Press (Belknap).
Pistor, Katherina. 1997. “Co-determination in Germany: A Socio-Political Model
with Governance Externalities.” (in this collection).
Prendergast, Canice. 1993. “The Role of Promotion in Inducing Specific Human
Capital Acquisition.” The Quarterly Journal of Economics 108 No. 2 (May): 523 534.
Putterman, Louis. 1984. “On Some Recent Explanations of Why Capital Hires
Labor.” Economic Inquiry 22 (April): 171-87.
Putterman, Louis, and Randall S. Kroszner. 1996. “The Economic Nature of the Firm:
A New Introduction.” In The Economic Nature of the Firm: A Reader (Second
Edition) edited by Putterman and Kroszner, 1-31. Cambridge University Press.
Rajan, Raghuram G., and Luigi Zingales. 1996. “Power in a Theory of the Firm.”
Working Paper.
Romano, Roberta. 1996. “Corporate Law and Corporate Governance,” paper prepared
for Conference on Firms, Markets and Organizations at University of California at
Berkeley, Haas School of Business, and scheduled for publication in a special issue of
Industrial and Corporate Change, Vol. 5, No. 3 (1996).
Rosen, Sherwin. 1985. “Implicit Contracts: A Survey.” Journal of Economic
Literature, Vol.
XXIII, September, pp. 1144-1175.
49
Shleifer, Andrei, and Lawrence H. Summers. 1988. “Breach of Trust in Hostile
Takeovers.” In Corporate Takeovers: Causes and Consequences, edited by Alan J.
Auerbach, 33-56. University of Chicago Press.
Topel, Robert C. 1990. “Specific Capital and Unemployment: Measuring the Costs
and Consequences of Job Loss.” In Studies in Labor Economics in Honor of Walter
Y. Oi, edited by Allan H. Meltzer and Charles I. Ploser, 181-214. Amsterdam: North
Holland.
________. 1991. “Specific Capital, Mobility, and Wages: Wages Rise with Job
Security.” Journal of Political Economy 99 (February): 145-176.
Vanek, Jaroslav. 1970. The General Theory of Labor-managed Market Economies.
Ithaca, NY: Cornell University Press.
------------. 1977. “The Basic Theory of Financing of Participatory Firms.” In The
Labor-Managed Economy: Essays by Jaroslav Vanek, edited by Vanek, 186-98.
Cornell University Press.
Ward, Benjamin. 1958. “The Firm in Illyria: Market Syndicalism,” American
Economic Review, Vol. 48, pp. 566-89.
Weiss, Andrew. 1990. Efficiency Wages: Models of Unemployment, Layoffs, and
Wage Dispersion. Princeton University Press.
Wiggins, Steven N. 1991. “The Economics of the Firm and Contracts: A Selective
Survey.” Journal of Institutional and Theoretical Economics 147 (December): 603-61.
Williamson, Oliver E. 1975. Markets and Hierarchies: Analysis and Antitrust
Implications. Free Press.
________. 1985. The Economic Institutions of Capitalism: Firms, Markets, Relational
Contracting. New York: The Free Press.
Willia